Professional Documents
Culture Documents
Received: 2020-09-30
Reviewed: 2020-10-12
Accepted: 2020-11-16
OnlineFirst: 2021-01-15
Published: 2021-04-01 DOI: https://doi.org/10.3916/C67-2021-05
Abstract
Research into risky online behaviour among children and adolescents is on the rise, with more studies being conducted
into the factors which can influence this phenomenon, above all in relation to school and family life. In the latter sphere,
one relevant factor is the degree of genuine trust children have in their parents when using the Internet. The main objective
of this study is to verify the effects of child disclosure about cyberaggression, in addition to the mediating role of problematic
Internet use and cybergossip, and the moderating role of gender and age. A total of 866 primary school children (53%
girls) between 10 and 13 years old (M=11.21; SD=0.90) were surveyed using self-reporting. The data processing followed
a moderated serial mediation model using “Process”. The results revealed the effects of child disclosure about
cyberaggression and the mediation of problematic Internet use and cybergossip. Unlike gender, age moderated the effects
of the mediation model. The results highlight the need to foster a climate of trust and communication in the family
environment to reduce involvement in risky online behaviour, in which children feel understood and supported by their
parents, which in turn encourages open communication about Internet use.
Resumen
La investigación sobre ciberconductas de riesgo infantil y juvenil se abre paso con estudios sobre factores que puedan
influir en estos fenómenos, entre los que se destacan los relacionados con la convivencia escolar y familiar. En esta última,
es relevante el nivel de confianza espontánea del hijo hacia su progenitor en el uso Internet. El objetivo de este trabajo es
comprobar el efecto de la revelación filial en la ciberagresión, así como el rol mediador del uso problemático de Internet y
el cibercotilleo, y el rol moderador del sexo y la edad. Un total de 866 escolares de primaria (53% chicas) de entre 10 y
13 años (M=11,21; D.T.=0,90) fueron encuestados mediante el uso de autoinformes. El tratamiento de datos siguió un
modelo de mediación serial moderada a través de «Process». Los resultados evidenciaron los efectos de la revelación
filial sobre la ciberagresión, así como la mediación del uso problemático de Internet y el cibercotilleo. A diferencia del
sexo, la edad moderó los efectos del modelo de mediación. Los resultados ponen de manifiesto la necesidad de establecer
un clima de confianza y comunicación en el entorno familiar para disminuir la implicación en las ciberconductas de riesgo,
donde los menores se sientan comprendidos y apoyados por los progenitores, facilitando la comunicación espontánea
sobre el uso de Internet.
The influence of the family context on children’s Internet behaviour has focused mainly on studying the norms
of control and supervision of the use of information and communication technologies (ICTs). These studies
show that parental control which imposes strict limits on children’s activities and continuously demands
information about what they are doing is in fact an ineffective strategy for reducing risky behaviour in
adolescence, including cyberaggression (Baldry et al., 2019; Sasson & Mesch, 2014). In particular, studies on
the forms of parental communication used to control their children’s online behaviour indicate that attempts to
obtain information through direct requests and questions and a certain level of intrusion is associated with
higher levels of cyberaggression (Shapka & Law, 2013). In contrast, the creation of a positive, trusting
relationship with both parents leads to better social adjustment and well-being in their children, and acts as a
protective factor, especially if child disclosure is encouraged, by which the family knows what the children are
doing because the children tell them, without the need for parental control (Machimbarrena et al., 2019). Here,
a number of studies have highlighted the fact that the risk of involvement in cyberaggression decreases with
the use of parental strategies based on communication and trust (Buelga et al., 2017).
Over 24% of young people use their digital devices on a daily basis to communicate with friends on social
media, and over half of these connect several times a day (Areepattamannil & Khine, 2017). However, it
constitutes problematic Internet use when it becomes compulsive and produces cognitive concern and a
deterioration of intrapersonal and interpersonal relations (Caplan, 2010) and stronger symptoms of depression
(Lozano-Blasco & Cortés-Pascual, 2020). Although it has been noted that personal factors can influence this
excessive use of technology, other studies indicate that the family plays a key role in preventing and reducing
it (Garmendia et al., 2019; Sela et al., 2020). Various studies have revealed how parental mediation, which is
Previous research has shown not only the negative consequences associated with being involved in
cyberaggression, an immoral practice in which a large number of adolescents, boys and girls alike, become
involved at an increasingly early age, but also the important role of the family in its prevention. However, further
research is needed looking into the family context in the development of this kind of online behaviour performed
by children and adolescents. Not only is it important to identify what family styles and strategies are most
effective in preventing online bullying, but also to explore how they can promote online behaviour that can help
reduce involvement in risky online behaviour and, therefore, foster positive online coexistence. In this context,
the present study aims to examine the mediating effect of problematic Internet use and cybergossip between
child disclosure and involvement in cyberaggression. Because a large number of studies have highlighted
which parental communication strategies, mainly those that enhance child disclosure, reduce the risk of
involvement in aggressive behaviour (Law et al., 2010), we expected to find that child disclosure about online
behaviour reduced cyberaggression (Hypothesis 1). Most previous work has focused on the effects of this
relationship, albeit without taking into account possible mediators. We also expected to find that problematic
Internet use and cybergossip mediated the relationship between child disclosure and cyberaggression
(Hypothesis 2). Due to the fact that boys tend to be more involved in risky online behaviour than girls throughout
adolescence, and that child disclosure is more common among girls and during childhood (Arpaci et al., 2020;
Smahel et al. al., 2020), we expected the effects of child disclosure on risky online behaviour to be lower in
late childhood (Hypothesis 3a) and in girls (Hypothesis 3b), compared to early adolescents and boys.
The incidental sample consisted of 866 schoolchildren (53% girls), from six state primary schools (58%) and
subsidized schools (42%), in both rural (36%) and urban (64%) areas in the provinces of Córdoba and Jaén
(Spain). The schools were selected through non-probability sampling for accessibility (Singleton & Straits,
2004). The children’s ages ranged from 10 to 13 years old (M=11.21; S.D.=0.90), with a distribution by primary
school year as follows: 28.8% in 4th year (n=249), 28.9% in 5th year (n=250) and 42.3% in 6th year (n=366).
Cyberaggression was assessed using 8 items on a five-point Likert-scale ranging from 0=never, 1=once or
twice, 2=once or twice a month, 3=about once a week, to 4=more than once a week. Examples of the items
for cyberaggression included: “I posted embarrassing videos or pictures of someone online” or “I spread
rumours about someone on the Internet” in the last two months. This scale was part of an adaptation for
primary school of the European Cyberbullying Intervention Project Questionnaire, ECIPQ (Del-Rey et al.,
2015), consisting of two dimensions, cybervictimization and cyberaggression. We used only the second
dimension for this study, taking the mean scores of the items, with higher values denoting a higher frequency
of being involved in cyberaggression. For this research, the validity indices were optimal: 𝜒2 S-B=56.4149;
gl=20; p<.001; NNFI=.98; CFI=.98; RMSEA=.04; SRMR=.05, and the reliability index was good for the
dimension of cyberaggression, ω = .74.
Child disclosure was measured with four items about how often the children told their parents about their online
activities, taken from the adaptation by Law et al. (2010) on Stattin and Kerr’s Parenting Questionnaire (Stattin
& Kerr, 2000). An ad hoc simple five-point Likert-type scale was used, ranging from “never” to “always”. The
following items were used: “I tell my parents when I open a new account on social media”, “I tell my parents
when I chat to new people on social media/WhatsApp”, “I tell my parents the content of the messages I send
or receive through social media/WhatsApp” and “I tell my parents before revealing information about myself to
other people through social media/WhatsApp”. Optimal indices were obtained for reliability (ω=.83) and
validity: 𝜒2 S-B=17.7493; gl=2; p<.001; NNFI=.98; CFI=.99; RMSEA=.06; SRMR=.02.
Problematic Internet use was measured using the Internet-Related Experiences Questionnaire scale (Beranuy
et al., 2009), which uses a 10-item Likert scale with 4 points of frequency, from “never” to “quite a lot”. The
items included the following: “I prefer to talk to my friends through social media/WhatsApp rather than in
person”, “I feel anxious when I cannot connect to social media/WhatsApp”. A unidimensional distribution of the
items was taken into account and optimal values were obtained with the study sample for reliability (ω=.78)
and validity: 𝜒2 S-B=51.3417; gl=35; p<.001; NNFI=.99; CFI=.99; RMSEA=.02; SRMR=.03.
Cybergossip was measured using the Cybergossip Questionnaire (Romera et al., 2018). This is a one-
dimensional survey consisting of nine items measured on a Likert scale with values from 0 (never) to 4
(always). The items included: “I tell my friends on social networks or WhatsApp about things I hear about that
happen to others”, “I talk about others on social networks or WhatsApp because it makes me feel closer to my
group of friends”. Optimal indices were obtained for the study sample for reliability (ω=.80) and validity, 𝜒2 S-
B=100.7370; gl=27; p<.001; NNFI=.96; CFI=.97; RMSEA=.05; SRMR=.05.
2.3. Procedure
The school management teams at the different schools were contacted to inform them about the aims of the
study. Those schools which expressed interest received detailed information on the data collection procedure.
Written consent was obtained from the students’ families and the participants gave their verbal consent. The
study complied with the Declaration of Helsinki on confidentiality, privacy and informed consent, and was
approved by the Ethical Committee for Bioethics and Biosafety of the University of Córdoba.
The questionnaires were administered individually on paper in the students’ classroom. The children were
informed that the test was voluntary and anonymous. Data collection was done by previously trained
researchers, who informed the children about the collection procedure and answered any questions they had
about how to complete the questionnaire. The teachers in charge of the class were absent during the process.
Response time ranged from 15 to 20 minutes. The data was collected in 2017.
Due to the skewness shown by the variables of cybergossip and cyberaggression, both were subjected to
square root transformation, following the recommendations by Tabachnick and Fidell (2007). All future
references refer to these transformed variables.
The preliminary analyses were performed using mean and standard deviation, together with the Spearman
correlation coefficient for each of the study factors which were to be verified in the theoretical model.
Similarly, Students’ t test was performed to contrast the existence of differences by gender and age (late
childhood=10-11 years old; early adolescence=12-13 years old), considering Cohen's d to control the effect
size. These analyses were performed using the IBM SPSS Statistics 20 statistical package.
3. Results
Table 1 shows the descriptive analyses of the study variables. Boys and early adolescents showed a greater
involvement in cyberaggression and problematic Internet use. Early adolescents also showed greater use of
cybergossip. Girls and late childhood participants showed higher levels of child disclosure. The effect size was
low-moderate. A positive relationship was observed between the variables of cyberaggression, cybergossip
and problematic Internet use, with scores ranging between .39 and .55. Child disclosure negatively correlated
with all the study variables (Table 1).
1. CA - .29 .35 .23 .28 2.51* .17 .20 .29 .33 .33 -6.17*** .42
2. CG .53 - .47 .40 .45 .38 .34 - .39 .36 .57 .41 -7.00*** .49
3. PIU .35 .47 - .84 .74 .66 .61 4.30*** .29 .67 .66 .85 .69 -3.46*** .24
4. CD -.25 -.24 -.24 2.26 1.36 2.68 1.25 -4.69*** .32 2.59 1.35 2.35 1.27 2.63*** .18
Key. M=mean; SD=Standard Deviation; t=Student’s t; d=Cohen’s d; CA=Cyberaggression; CC=Cybergossip; PIU=Problematic Internet
Use; CD=Child disclosure; *p<.05; **p<.01; ***p<.001. 1 1=Boys; 2=Girls. 2 1=Late childhood; 2=Ealy adolescence.
A serial mediation analysis was carried out with 5,000 bootstrap samples. Table 2 shows the coefficients of
the mediation model. The total effect of child disclosure on cyberaggression was significant: β=-.27, t=-8.15,
p<.001. We then analysed the models of the mediating variables, problematic Internet use and cybergossip,
and of the dependent variable, cyberaggression. Child disclosure was significantly associated with problematic
internet use and cybergossip. Problematic internet use was significantly associated with cybergossip and
cyberaggression. Cybergossip was also significantly associated with cyberaggression. After controlling for the
effects of mediators, the direct effect of child disclosure on cyberaggression remained significant. The
bootstrap procedure was used to evaluate the indirect effect and the confidence intervals (CI). An indirect
effect was significant if the CI did not include the value 0. In addition, a significant indirect effect was obtained
for the Child disclosure → Problematic Internet use → Cyberaggression path, β=-.04, 95% CI [-.06, -.02],
Pm=.14. The Child disclosure → Cybergossip → Cyberaggression path produced a significant indirect effect,
β=-.05, 95% CI [-.09, -.02], Pm=.20. Finally, the Child disclosure → Problematic Internet use → Cybergossip
→ Cyberaggression path showed a significant indirect effect, β=-.15, 95% CI [-.19, -.10], Pm=.55. Thus,
problematic internet use and cybergossip were shown to have a partial mediating role in the effect of child
disclosure on cyberaggression.
The moderation results indicate that there were no statistically significant differences between boys and girls
in the associations between variables (ps>.05). Age proved to be a determining variable in the effects of child
disclosure on problematic Internet use: ß=-.20, t=-2,95 p<.01. The effects were significantly greater in early
adolescence in comparison to late childhood (See Figure 1).
Figure 1. Moderating role of age in the effects of child disclosure on problematic Internet use
Age was found to influence the effects of problematic Internet use on cybergossip: β=.21, t= .51, p<.001. These
effects were significantly greater in early adolescence in comparison to late childhood (Figure 2). Indirect
effects show how the Child disclosure → Problematic Internet Use → Cyberaggression path was significant in
early adolescents, βEarly-adolescence= .07, 95% CI [-.12, -.03], but not in late childhood, βlate childhood=-0.02, 95% CI
[-.04, .00]. As regards the Child disclosure → Cybergossip → Cyberaggression path, there were no significant
differences between the two groups, βEarly adolescence – late childhood=.00, 95% CI [-.06, .06]. Finally, as for the Child
disclosure → Problematic Internet use → Cybergossip → Cyberaggression path, early adolescents showed
statistically greater effects, βEarly adolescence=-.09, 95% CI [-.13, -.06], βlate childhood=-.02, 95% CI [-.04, -.01], βEarly
adolescence – late childhood=-.06, 95% CI [-.11, -.03].
Figure 2. Moderating role of age in the effects of problematic Internet use on cybergossip
References
Álvarez-García, D., Núñez, J.C., González-Castro, P., Rodríguez, C., & Cerezo, R. (2019). The effect of parental control
on cyber-cictimization in adolescence: The mediating role of impulsivity and high-risk behaviors. Frontiers in
Psychology, 10, 1-7. https://doi.org/10.3389/fpsyg.2019.01159
Areepattamannil, S., & Khine, M.S. (2017). Early adolescents’ use of information and communication technologies (ICTs)
for social communication in 20 countries: Examining the roles of ICT-related behavioral and motivational
characteristics. Computers in Human Behavior, 73, 263-272. https://doi.org/10.1016/j.chb.2017.03.058
Arpaci, I., Abdeljawad, T., Baloğlu, M., Kesici, Ş., & Mahariq, I. (2020). Mediating effect of Internet addiction on the
relationship between individualism and cyberbullying: Cross-sectional questionnaire study. Journal of Medical
Internet Research, 22(5), 1-24. https://doi.org/10.2196/16210
Baldry, A.C., Sorrentino, A., & Farrington, D.P. (2019). Cyberbullying and cybervictimization versus parental supervision,
monitoring and control of adolescents’ online activities. Children and Youth Services Review, 96, 302-307.
https://doi.org/10.1016/j.childyouth.2018.11.058
Bartau-Rojas, I., Aierbe-Barandiaran, A., & Oregui-González, E. (2018). Parental mediation of the Internet use of Primary
students: Beliefs, strategies and difficulties. [Mediación parental del uso de Internet en el alumnado de Primaria:
creencias, estrategias y dificultades]. Comunicar, 54, 71-79. https://doi.org/10.3916/C54-2018-07
Beranuy, M., Chamarro, A., Graner, C., & Carbonell-Sánchez, X. (2009). Validación de dos escalas breves para evaluar
la adicción a Internet y el abuso del móvil. Psicothema, 21, 480-485.
Buelga, S., Martínez-Ferrer, B., & Cava, M.J. (2017). Differences in family climate and family communication among
cyberbullies, cybervictims, and cyber bully-victims in adolescents. Computers in Human Behavior, 76, 164-173.
https://doi.org/10.1016/j.chb.2017.07.017
Caplan, S.E. (2010). Theory and measurement of generalized problematic Internet use: A two-step approach. Computers
in Human Behavior, 26(5), 1089-1097. https://doi.org/10.1016/j.chb.2010.03.012
Cerezo, F., Arnaiz, P., Giménez, A.M., & Maquilón, J.J. (2016). Online addiction behaviors and cyberbullying among
adolescents. Anales de Psicologia, 32(3), 761-769. https://doi.org/10.6018/analesps.32.3.217461
Chen, L., & Shi, J. (2019). Reducing harm from media: A meta-analysis of parental mediation. Journalism & Mass
Communication Quarterly, 96(1), 173-193. https://doi.org/10.1177/1077699018754908
Chernick, M.R. (2008). Bootstrap methods: A guide for practitioners and researchers. Wiley.
https://doi.org/10.1002/9780470192573
Del-Rey, R., Casas, J.A., Ortega-Ruiz, R., Schultze-Krumbholz, A., Scheithauer, H., Smith, P., Thompson, F., Barkoukis,
V., Tsorbatzoudis, H., Brighi, A., Guarini, A., Pyzalski, J., & Plichta, P. (2015). Structural validation and cross-cultural
robustness of the European Cyberbullying Intervention Project Questionnaire. Computers in Human Behavior, 50,